Historical data reveal contrasting habitat amount relationships with plant biodiversity

Assessing habitat loss effects on biodiversity is a major focus of ecological research. The relationship between habitat amount and biodiversity, postulated in the habitat amount hypothesis, is usually assessed at one point in time, which does not account for habitat loss as a temporal process. We examined habitat amount effects at two time periods, 1930s and 2010s, using plant data from three semi-natural habitats: calcareous grassland, heathland and broadleaved woodland, across Dorset, southern England. Woodlands, which changed little in area over the time period, showed minimal effects of habitat amount on species occurrence in both time periods. For grassland and heathland, which had undergone severe losses over the study period, we found the expected positive relationship in the 2010s, but the relationship was negative for these habitats in the 1930s. We explored possible reasons for this result. Total perimeter-to-area ratio (TPAR) showed positive effects in the 1930s for grassland and heathland, suggesting effects of habitat configuration, specifically edge. However, TPAR was highly correlated with habitat amount so this finding is speculative. One possible explanation for the relationships with habitat amount, and the change between the two periods could be the quality of the surrounding matrix. In the 1930s, the landscape was less intensified and was dominated by semi-natural habitats, whereas by the 2010s much had been converted to arable and intensive grasslands. We speculate that species could likely utilise the matrix to a greater degree in the 1930s compared with the 2010s when the matrix was more hostile, thereby decreasing the importance of habitat amount in the 1930s compared with the 2010s. These findings have important implications for conservation, as they show the importance of context (i.e. matrix quality) in determining the relationship between habitat amount and biodiversity